APIVAI API Examples

APIVAI is an OpenAI- and Anthropic-compatible API gateway for Claude and GPT models at a fraction of official list price. Because it's compatible, existing tools (Claude Code, Cursor, Cline, Codex CLI, the OpenAI/Anthropic SDKs) work by changing the base URL and key — no custom code. Pay-as-you-go, with crypto / USDT / Alipay, no VPN, no subscription.

Developer-focused examples for calling APIVAI with OpenAI-compatible request patterns. The repository keeps each example intentionally small so you can inspect the request shape, confirm credentials, and adapt the code to your own application.

• APIVAI website: https://apivai.com/
• Product documentation: https://apivai.com/docs
• OpenAI-compatible API guide: https://apivai.com/openai-compatible-api
• Claude API proxy notes: https://apivai.com/claude-api-proxy
• Pricing: https://apivai.com/pricing

Default API base URL used by these examples:

https://api.apivai.com/v1

What this repository is for

Use this repository when you want to:

• Check that your APIVAI API key and network path are configured correctly.
• See minimal Python, Node.js, and cURL requests before integrating a full SDK or application.
• Verify model names with GET /v1/models instead of hardcoding a value.
• Keep example code simple enough to copy into a local prototype and then replace with your own error handling, logging, and configuration.

This repository is not a production application framework. Treat it as a set of small connectivity and request-format examples.

Example entry points

Runtime	File	Purpose
Python	examples/python-chat.py	Sends a chat completion request from a small Python script.
Node.js	examples/node-chat.js	Sends a chat completion request with the built-in Node.js fetch API.
cURL	examples/curl-chat.md	Shows a shell-based request that is useful for quick endpoint testing.
Python streaming	examples/python-streaming.py	Reads streamed chat-completion chunks from APIVAI.
Node.js streaming	examples/node-streaming.js	Reads streamed chat-completion chunks with built-in Node.js APIs.
Models cURL	examples/models-list-curl.md	Lists available models before setting APIVAI_MODEL.

SDK & framework examples

Because the endpoint is OpenAI- and Anthropic-compatible, the official SDKs and popular frameworks work by overriding the base URL and key — no custom integration code.

SDK / framework	File	Notes
OpenAI SDK (Python)	examples/openai-sdk-python.py	Official openai package, base_url override.
OpenAI SDK (Node)	examples/openai-sdk-node.js	Official openai package, baseURL override.
Anthropic SDK (Python)	examples/anthropic-sdk-python.py	Official anthropic package against /v1/messages.
Anthropic SDK (Node)	examples/anthropic-sdk-node.js	Official @anthropic-ai/sdk package, baseURL override.
LangChain (Python)	examples/langchain-python.py	ChatOpenAI(base_url=...).
LlamaIndex (Python)	examples/llamaindex-python.py	OpenAILike(api_base=...).
Function calling (Python)	examples/function-calling-python.py	Tool-use round-trip via the OpenAI SDK.

Use with your editor or agent

These tools accept an OpenAI-compatible base URL + key — set https://api.apivai.com/v1 and a model from GET /v1/models:

Tool	Guide
Claude Code	docs/claude-code.md
Cursor	docs/cursor.md
Codex CLI	docs/codex-cli.md
Cline	docs/cline.md
Aider	docs/aider.md
Continue.dev	docs/continue-dev.md

For Claude-specific workflows, confirm whether your tool expects OpenAI-compatible or Anthropic-compatible settings. See the APIVAI Claude API proxy page for current guidance: https://apivai.com/claude-api-proxy.

Requirements

Before running examples, make sure you have:

• An APIVAI API key from your account.
• Network access to https://api.apivai.com/v1.
• One of the following local runtimes:
  • Python 3.9+ for examples/python-chat.py.
  • Node.js 18+ for examples/node-chat.js.
  • curl for examples/curl-chat.md.

Quick start

1) Set environment variables

All examples read these environment variables:

• APIVAI_API_KEY: your secret API key.
• APIVAI_BASE_URL: defaults to https://api.apivai.com/v1 when not set in supported examples.
• APIVAI_MODEL: a model name returned for your account.

macOS/Linux:

export APIVAI_API_KEY="YOUR_APIVAI_API_KEY"
export APIVAI_BASE_URL="https://api.apivai.com/v1"
export APIVAI_MODEL="YOUR_MODEL_NAME"

PowerShell:

$env:APIVAI_API_KEY="YOUR_APIVAI_API_KEY"
$env:APIVAI_BASE_URL="https://api.apivai.com/v1"
$env:APIVAI_MODEL="YOUR_MODEL_NAME"

2) Discover available models

Model availability can vary by account and over time. Start by calling:

GET https://api.apivai.com/v1/models

Example:

curl -s "$APIVAI_BASE_URL/models" \
  -H "Authorization: Bearer $APIVAI_API_KEY" \
  -H "Content-Type: application/json"

Set APIVAI_MODEL to a model name returned by that response.

3) Run an example

python3 examples/python-chat.py

node examples/node-chat.js

For cURL, follow examples/curl-chat.md.

Common errors / troubleshooting

• 401 Unauthorized: Confirm that APIVAI_API_KEY is set, has no extra spaces, and is sent as Authorization: Bearer <key>.
• 404 Not Found: Check that APIVAI_BASE_URL includes /v1 and that the endpoint path is correct.
• model_not_found or similar model errors: call GET /v1/models again and use a returned model name.
• 429 Too Many Requests: Reduce request rate and retry with exponential backoff.
• 5xx responses: retry after a short delay and capture any request identifiers or timestamps for support/debugging.
• Empty or malformed responses: verify that the request body is valid JSON and that Content-Type: application/json is present.

For a longer setup walkthrough, see docs/setup.md. For testing commands, see docs/testing.md. For common questions, see docs/faq.md.

Additional developer guides:

• docs/openai-compatible-endpoint.md explains the OpenAI-compatible base URL, request shape, and streaming behavior.
• docs/claude-code.md maps APIVAI environment variables to Claude Code-style OpenAI-compatible settings.
• docs/cursor.md outlines editor configuration checks for Cursor-style tools.
• docs/troubleshooting.md provides status-code and streaming debugging steps.

Security notes

• Never commit real API keys, tokens, or account identifiers.
• Do not expose secret keys in browser or mobile client code; route requests through a server you control.
• Prefer environment variables, a local .env file excluded from Git, or a managed secrets store.
• Rotate keys immediately if they are pasted into logs, screenshots, issues, pull requests, or chat messages.
• Review logs before sharing them publicly; request and response bodies may contain sensitive user data.

Repository structure

.
├── README.md
├── .env.example
├── examples/
│   ├── python-chat.py
│   ├── node-chat.js
│   ├── curl-chat.md
│   ├── python-streaming.py
│   ├── node-streaming.js
│   ├── models-list-curl.md
│   ├── openai-sdk-python.py
│   ├── openai-sdk-node.js
│   ├── anthropic-sdk-python.py
│   ├── langchain-python.py
│   ├── llamaindex-python.py
│   └── function-calling-python.py
├── docs/
│   ├── setup.md
│   ├── faq.md
│   ├── testing.md
│   ├── openai-compatible-endpoint.md
│   ├── claude-code.md
│   ├── cursor.md
│   ├── codex-cli.md
│   ├── cline.md
│   ├── aider.md
│   └── troubleshooting.md
└── .github/
    └── ISSUE_TEMPLATE/

License

See LICENSE.